Target material for deposition of molybdenum oxide layers

Abstract

An electrically conductive, oxidic target material includes a proportion of substoichiometric molybdenum oxide phases of at least 60% by volume, an MoO.sub.2 phase in a proportion of 2-20% by volume, and optionally an MoO.sub.3 phase in a proportion of 0-20% by volume. The substoichiometric molybdenum oxide phase proportion is formed by one or more substoichiometric MoO.sub.3-y phase(s), where y is in each case in a range from 0.05 to 0.25. A process for producing the target material and a process for using the target material are also provided.

Claims

1. An electrically conductive, oxidic target material configured for sputtering, comprising: a proportion of substoichiometric molybdenum oxide phases of at least 60% by volume, based upon a total volume of solid material in the electrically conductive, oxidic target material comprising substoichiometric MoO.sub.3-y phases, where y is in each case in a range of from 0.05 to 0.25; an MoO.sub.2 phase in a proportion of 2-20% by volume, based upon the total volume of solid material in the electrically conductive, oxidic target material; and a proportion of an MoO.sub.3 phase being 1% by volume, based upon the total volume of solid material in the electrically conductive, oxidic target material, wherein the electrically conductive, oxidic target material is a sintered material having a relative density of at least 95%, wherein the electrically conductive, oxidic target material further comprises a metal other than molybdenum present in metallic or oxidic form as a dopant in an amount of from 0.5 mol % to 20 mol %, wherein the dopant is a metal selected from the group consisting of tantalum, niobium, titanium, chromium, zirconium, vanadium, hafnium, and tungsten, and wherein the electrically conductive, oxidic target material has an oxygen content in a range of from 71.4 to 74.5 at. %.

2. The electrically conductive, oxidic target material according to claim 1, wherein the proportion of the substoichiometric molybdenum oxide phases is at least 85% by volume, based upon the total volume of solid material in the electrically conductive, oxidic target material, and the proportion of the MoO.sub.2 phase is in a range of 2-15% by volume.

3. The electrically conductive, oxidic target material according to claim 1, wherein the proportion of substoichiometric molybdenum oxide phases is formed by at least one substoichiometric phase MO.sub.4O.sub.11, MO.sub.17O.sub.47, MO.sub.5O.sub.14, MO.sub.8O.sub.23, MO.sub.9O.sub.26 or MO.sub.18O.sub.52.

4. The electrically conductive, oxidic target material according to claim 3, wherein the substoichiometric phase Mo.sub.4O.sub.11 is at least 45% by volume, based upon the total volume of solid material in the electrically conductive, oxidic target material.

5. The electrically conductive, oxidic target material according to claim 1, wherein the dopant is tantalum.

6. The electrically conductive, oxidic target material according to claim 1, wherein the dopant is niobium.

7. The electrically conductive, oxidic target material according to claim 1, wherein the electrically conductive, oxidic target material has a relative density of >98%.

8. A process for producing the electrically conductive, oxidic target material according to claim 1, the process comprising the following steps: providing a molybdenum oxide-containing powder or a molybdenum oxide-containing powder mixture having an oxygen content matched to the electrically conductive, oxidic target material to be produced; introducing the powder mixture into a mold; and densifying the powder mixture by pressure, heat, or pressure and heat.

9. The process according to claim 8, which further comprises carrying out the densifying step by hot pressing, hot isostatic pressing, spark plasma sintering or pressing-sintering.

10. A process for using the electrically conductive, oxidic target material according to claim 1, the process comprising the following step: carrying out a sputtering process as a DC sputtering process or a pulsed DC sputtering process in a noble gas atmosphere without oxygen or with an introduction of not more than 20% by volume of oxygen as a reactive gas.

Description

BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWING

(1) FIG. 1: Binary phase diagram of the molybdenum-oxygen system; source: Brewer L., and Lamoreaux R. H. in Binary Alloy Phase Diagrams, II Ed., Ed. T. B. Massalski, 1990.

(2) FIG. 2: Raman reference spectrum 1 (MoO.sub.2), where the intensity (counts) is plotted against the Raman shift (cm.sup.1).

(3) FIG. 3: Raman reference spectrum 2 (Mo.sub.4O.sub.11).

(4) FIG. 4: Raman reference spectrum 3 (presumably Mo.sub.18O.sub.52).

(5) FIG. 5: Raman reference spectrum 4 (substoichiometric molybdenum oxide having an unknown composition, but not Mo.sub.6O.sub.14, Mo.sub.8O.sub.23 or Mo.sub.9O.sub.26).

(6) FIG. 6: Raman reference spectrum 5 (MoO.sub.3).

(7) FIG. 7: Raman reference spectrum 6 (Ta.sub.2O.sub.5).

(8) FIG. 8: A microstructure of a third working example produced by means of Raman mapping.

(9) FIG. 9: A microstructure of a fourth working example produced by means of Raman mapping.

(10) FIG. 10: A microstructure of a fifth working example produced by means of Raman mapping.

(11) FIG. 11: Reflection properties of a molybdenum-tantalum oxide layer deposited under different process parameters (sputtering power, process gas pressure).

WORKING EXAMPLES

Example 1

(12) MoO.sub.3 powder (Molymet) having an average particle size of 4.4 m is reduced at 550 C. in an H.sub.2 atmosphere (dew point (H.sub.2)=10 C.) for 17 minutes in a furnace. The Mo oxide powder obtained has an oxygen content of 73.1 at. %. It is placed in a graphite mould having the dimensions 260240 mm and a height of 50 mm and densified in a hot press under vacuum at a pressing pressure of 45 MPa, a temperature of 750 C. and a hold time of 120 minutes. The compacted component displays a relative density (pore determination on a metallographic polished section) of 96% and comprises an MoO.sub.2 phase in a proportion of 10% by volume, an MoO.sub.3 phase in a proportion of 7% by volume and a proportion of substoichiometric molybdenum oxide phases of 83% by volume. The substoichiometric molybdenum oxide phase component is formed predominantly by Mo.sub.4O.sub.11. The determination of the phase composition in this and the following examples is carried out by means of Raman mapping and is explained in detail at the end of the examples.

Example 2

(13) 36.2 mol % of MoO.sub.2 powder (Plansee) and 63.8 mol % of MoO.sub.3 powder (Molymet) are mixed and homogenized for 30 minutes in a ball mill equipped with zirconium oxide mixing balls (diameter 10 mm). The resulting powder mixture having an oxygen content of 72.5 at. % is placed in a graphite mould having a diameter of 70 mm and a height of 50 mm and densified in a spark plasma sintering (SPS) plant under vacuum at a pressing pressure of 40 MPa, a temperature of 775 C. and a hold time of 120 minutes. The compacted component has a relative density of 98%. It consists of an MoO.sub.2 phase in a proportion of 2.7% by volume and a proportion of substoichiometric molybdenum oxide phases of 97.3% by volume in total. An MoO.sub.3 phase could not be detected. The substoichiometric molybdenum oxide phase component is formed to an extent of 53% by volume by Mo.sub.4O.sub.11.

Example 3

(14) The proportion of coarse particles and agglomerates is sieved out from MoO.sub.2 powder (Plansee SE) in a sieve (mesh opening 32 m). 24 mol % of the MoO.sub.2 powder obtained are mixed with 70 mol % of MoO.sub.3 powder (Molymet) and 6 mol % of tantalum powder in a ploughshare mixer (Lodige) for 20 minutes so as to obtain a homogeneous distribution between the powder components. The powder mixture obtained is placed in a graphite mould having the dimensions 260240 mm and a height of 50 mm and densified in a hot press under vacuum at a pressing pressure of 40 MPa, a temperature of 750 C. and a hold time of 60 minutes. The compacted component has a relative density of 95.6%. The target material obtained comprises an MoO.sub.2 phase in a proportion of 10.3% by volume, an MoO.sub.3 phase in a proportion of 19.2% by volume, substoichiometric molybdenum oxides in a total proportion of 68.4% by volume and a Ta.sub.2O.sub.5 phase in a proportion of 2.1% by volume. The predominant component of the substoichiometric molybdenum oxides is Mo.sub.4O.sub.11 in a proportion of 49.4% by volume. The further substoichiometric molybdenum oxides are (presumably) Mo.sub.18O.sub.52 and a not yet known substoichiometric molybdenum oxide having an unknown composition. The Raman spectrum of this substoichiometric molybdenum oxide is shown in FIG. 5. The microstructure of the target material determined by means of Raman mapping is shown in FIG. 8. Regions comprising MoO.sub.2 phase, regions comprising MoO.sub.3 phase and regions comprising Ta.sub.2O.sub.5 phase are discernible in the microstructure; these various phases are embedded as islands in a contiguous network formed by the substoichiometric molybdenum oxides Mo.sub.4O.sub.11, Mo.sub.18O.sub.52 and the substoichiometric molybdenum oxide having an unknown composition.

Example 4

(15) Example 4 differs from Example 3 by variation of the hot pressing parameters; manufacture of the powder batch is carried out as in Example 3. The powder mixture is placed in a graphite mould having the dimensions 260240 mm and a height of 50 mm and densified in a hot press under vacuum at a pressing pressure of 40 MPa, a temperature of 750 C. and a hold time of 240 minutes. The compacted component has a relative density of 97%. The target material obtained comprises an MoO.sub.2 phase in a proportion of 8.1% by volume, an MoO.sub.3 phase in a proportion of 5.5% by volume, substoichiometric molybdenum oxides in a total proportion of 85% by volume and a Ta.sub.2O.sub.5 phase in a proportion of 1.4% by volume. The predominant component of the substoichiometric molybdenum oxides is Mo.sub.4O.sub.11 in a proportion of 59.1% by volume. FIG. 9 shows the microstructure of the target material produced by means of Raman mapping.

Example 5

(16) Example 5 differs from Examples 3 and 4 by variation of the hot pressing parameters; the manufacture of the powder batch is carried out as in Example 3. The powder mixture is placed in a graphite mould having the dimensions 260240 mm and a height of 50 mm and densified in a hot press under vacuum at a pressing pressure of 40 MPa, a temperature of 790 C. and a hold time of 120 minutes. The compacted component has a relative density of 99.7%. The target material obtained comprises an MoO.sub.2 phase in a proportion of 5.7% by volume, substoichiometric molybdenum oxides in a proportion of 91.9% by volume and Ta.sub.2O.sub.5 phase in a proportion of 2.4% by volume. An MoO.sub.3 phase is not detectable. Among the substoichiometric molybdenum oxides, Mo.sub.4O.sub.11 with a proportion of 47.2% by volume and the substoichiometric molybdenum oxide having an unknown composition with a proportion of 31.4% by volume make up the largest part. The Raman spectrum of the still unknown substoichiometric molybdenum oxide is depicted in FIG. 5. The microstructure of the target material produced by means of Raman mapping is shown in FIG. 10.

(17) To determine the proportions by volume of the various phases and the density of the target material, a metallographic polished section was produced by means of dry preparation from a representative part of a specimen by cutting a specimen having an area of about 10-1510-15 mm.sup.2 to size in a dry cutting process (diamond wire saw, bandsaw, etc.), cleaning it by means of compressed air, subsequently embedding it hot and conductively (C-doped) in phenolic resin, grinding and polishing it. Since at least the MoO.sub.3 phase proportion is water-soluble, dry preparation is important. The polished section obtained in this way was subsequently analysed under an optical microscope.

(18) For the positionally resolved determination of the molybdenum oxide phases, use was made of a Raman microscope (Horiba LabRAM HR800) in which a confocal optical microscope (Olympus BX41) is coupled with a Raman spectrometer. The surface to be analysed was scanned over an area of 11 mm.sup.2 by means of a focus laser beam (HeNe laser, wavelength =632.81 nm, 15 mW total power) point-by-point in steps of 10 m (the sample surface to be examined was fixed on a motorized XYZ table and the latter was moved). A complete Raman spectrum was produced for each one of the 100100 measurement points (Raman mapping). Raman spectra are obtained from the backscattered radiation and are wavelength-dispersively split up by means of an optical grating (300 lines/mm; spectral resolution: 2.6 cm.sup.1) and recorded by means of a CCD detector (1024256 pixel multichannel CCD; spectral range: 200-1050 nm). In the case of a microscope objective having 10 enlargement and a numerical aperture NA of 0.25, which serves for focussing the laser beam from the Raman spectrometer, it was possible to achieve a theoretical measurement point size of 5.2 m.sup.2. The excitation energy density (3 mW/m.sup.2) is selected low enough to avoid phase transformations in the specimen. The penetration depth of the excitation radiation is limited to a few microns in the case of molybdenum oxides (in the case of pure MoO.sub.3 here about 4 m; but since a mixture of different phases is analysed, precise indication of the penetration depth is not possible). For each measurement point, the Raman signal was averaged over an acquisition time of 4 s, which gave a sufficiently good signal-to-noise ratio. A two-dimensional depiction of the surface composition of the specimen was produced by automated evaluation of these Raman spectra (evaluation software Horiba LabSpec 6) and the domain size, proportions by area, etc., of the various phases can be determined quantitatively therefrom. For precise identification of a molybdenum oxide phase, reference spectra of previously synthesized reference specimens or reference specimens of relatively large homogeneous specimen regions are recorded, with it being ensured that a reference spectrum corresponds precisely to one metal oxide phase. In FIGS. 2 to 7 typical reference spectra of MoO.sub.2, MO.sub.4O.sub.11, Mo.sub.18O.sub.52, a previously unknown substoichiometric MoO.sub.x oxide, MoO.sub.3 and Ta.sub.2O.sub.5 are shown (the intensity (count) of the scattered light versus the Raman shift (cm.sup.1) is shown in the individual spectra). The analysis and assignment of the Raman spectra is carried out using the Multivariant Analysis Modules of the abovementioned evaluation software by means of the CLS method (classical least squares fitting). The specimen spectrum S is for this purpose represented as a linear combination of the individual normalized reference spectra R.sub.i, where c.sub.i is the respective weighting factor and is an offset value, S=c.sub.iR.sub.i+. A colour corresponding to a metal oxide phase is subsequently assigned to each measurement point, with only the phase having the greatest weighting factor c.sub.i being used for the colour assignment in each case. The magnitude (the absolute value) of the weighting factor c.sub.i determines the brightness of the measurement point. This procedure is justified since the spectrum of one measurement point can generally be unambiguously assigned to a single metal oxide phase.

(19) For the objective used, a specimen spectra was obtained from all 100100 measurement points, even when the measurement was made on a pore. In this case, the signal originated from a lower region located under the pore. If no Raman spectrum is obtained for individual measurement points, e.g. owing to a pore, this can be excluded from the determination of the proportions by area, i.e. the volume occupied by the pores of the target material is excluded from the total volume. The reported volumes for the individual molybdenum oxide phases therefore on their own add up to 100% without the pore volume.

(20) The method of analysis described here is particularly suitable for determining the relative proportions of the phase of various Mo oxides. In a repeat measurement (one specimen was measured 3 times in succession), a relative measurement error of 10% (based on the phase component determined in each case) was found. The relative measurement error in the % by volume determination of the dopant oxides (e.g. Ta.sub.2O.sub.5) on the other hand was 25%. It is therefore possible that the measured percentages by volume in the examples deviate slightly from the weighed-out amounts of dopant metal (e.g. tantalum) or dopant metal oxide (e.g. Ta.sub.2O.sub.5).

(21) The determination of the relative density is carried out by means of digital image analysis of optical micrographs of the metallographic polished section, in which the relative proportion by area of the pores is determined. For this purpose, preparation of the specimens was followed by recording of in each case three bright field micrographs having a size of 11 mm with 100 enlargement, with zones of obvious cavities or other damage such as scratches caused by dry preparation being avoided where possible. The images obtained were evaluated by means of the digital image analysis software implemented in the IMAGIC image data bank. For this purpose, the pore component (dark) was marked on the image as a function of the grey scale by means of a histogram. The lower limit of the interval was set at 0 (=black). On the other hand, the upper limit has to be estimated subjectively with the aid of the grey scale intensity histogram (255=white). The image region to be measured was set (ROI) in order to exclude the scale bars. The relative proportion by area (in percent) and the image coloured according to the selected grey scale interval (coloured means that this pixel was included in the measurement and accordingly counted as pore) is obtained as result. The value for the relative density was determined as arithmetic mean of three such porosity measurements.

Example 6

(22) In a series of experiments, the molybdenum-tantalum oxide target produced as described in Example 3 was nonreactively sputtered under different process conditions in order to check the reproducibility and process stability by means of the properties of the layer. Here, the reflectivity of the layers produced was used as criterion for the assessment. To determine the reflectivity, glass substrates (Corning Eagle XG, 50500.7 mm.sup.3) were coated with molybdenum-tantalum oxide and a covering layer of 200 nm of Al. The reflection was measured through the glass substrate using a Perkin Elmer Lambda 950 photospectrometer. In order to obtain a very low reflectivity, the layer thickness of the molybdenum oxide was varied in the range from 40 to 60 nm in a first experiment, with the best result being able to be achieved at a layer thickness of 51 nm. This layer thickness was subsequently used and kept constant for all further experiments. Results from this series of experiments are shown in FIG. 11. In the experiments, the sputtering power was varied in the range from 400 W to 800 W and the process pressure of the argon was varied in the range from 2.510.sup.3 mbar (11 sccm) to 1.010.sup.2 mbar (47 sccm). It can be seen that both the increase in the power by a factor of 2 from 400 W to 800 W and also the increase in the process pressure by a factor of 2 from 510.sup.3 mbar to 110.sup.2 mbar have only a negligible influence on the measured properties of the layer.

(23) The high process stability of the sputtering process using a molybdenum-tantalum oxide target was thus confirmed. Reproducible results can be achieved in a wide process window, in complete contrast to the reactive sputtering process using a metallic target, which is highly unstable.